Beilstein J. Org. Chem. 2020, 16, 1588–1595.
Next we used NMR spectroscopy to further characterize the
binding behavior. Initial titrations of YY or FF to a(CFC)2 or
p(CFC)2 hinted at formation of H-bonds due to shifts of
protons at α-carbon atoms between δ = 3.85 to 4.00. However,
no binding constants could be determined from those measure-
ments as even in a 10-fold excess the system showed no signs
behavior does not contradict the ITC measurements and can be
attributed to the kinetic isotope effect. Although, D-bonding is
generally stronger than H-bonding, its contribution to the
peptide–peptide binding can be weakened by even stronger
competitive binding to the D2O solvent [24]. The shifting
signals could be used to obtain Job plots which indicate 1-to-1
complexes for all combinations of YY or FF with a(CFC)2 or
x ≈ 0.5. Their profiles show shallow curves instead of sharp
transitions, which is in agreement with the apparently weaker
Conclusion
Figure 3: ITC of YY (30 mM) to a(CFC)2 (1.5 mM) in phosphate buffer
(pH 7.4, 100 mM).
In summary, we demonstrated that artificial peptide receptors
can emerge from a peptide-based DCL under competitive
conditions in water. In agreement with our previous works on
In the interaction of FF with a(CFC)2 and p(CFC)2 the carbohydrate recognition, this supports our initial assumption
enthalpy is very low, which is cause for a poor signal-to-noise that dynamic peptides, due to their similarity towards enzymes,
ratio. Other values for these particular interactions should there- can bind a broad scope of biomolecules. ITC measurements
fore be met with some caution. FF seems to bind stronger to suggest that the cyclic tripeptide dimers a(CFC)2 and p(CFC)2
a(CFC)2 than to p(CFC)2, though again the values show some are stronger binders for the aromatic dipeptides FF and YY
uncertainty and real values of the binding constants are likely in (K ≈ 229–702 M−1), then for the non aromatic AA, which is
the same order of magnitude as for YY. Unlike for the interac- bound less efficiently with a factor of about 3–10 times weaker
tion with FF, the entropy change of interaction with YY (K ≈ 65–71 M−1).
appears to be highly positive, which might be due to hydro-
Experimental
phobic interactions. We do not attribute this behavior to a
fundamentally different mode interaction compared to FF Preparation of the peptide libraries
though, as both dipeptides behave very similar in the following Stock solutions of each CXC building blocks (7.5 mM each)
NMR experiments. The effect may be attributed to the intermo- were prepared in NH4CO3 buffer (pH 7.8, 100 mM). 50 µL
lecular interaction of the dipeptides with themselves. For stock solution of each was given to a 1.5 mL vial. Another
example, FF is a popular self-assembly motif, which also 100 µL of a template stock solution (90 mM) was added. For a
explains its poor solubility (≈5 mM) [23]. Hence the thermo- reference sample the template solution was substituted for
dynamic data depicts not only the association of a(CFC)2/ buffer. Small magnetic stirring bars were added, and the vials
p(CFC)2 towards FF, but also the dissociation of FF were covered by a layer of perforated Parafilm to slow down
assemblies. The aromatic side groups of FF and YY seem to drying due to fast air exchange. After stirring overnight
play an important role in the binding, since AA binds signifi- (100 rpm), the samples were acidified with a 5% solution of
cantly weaker. Each of the aromatic units contributes to the formic acid in water. The samples were diluted 1:10 with water
binding to a roughly equal amount, as is shown by a compari- and measured with HPLC–MS setup I.
son with AF and FA (Scheme 2c), which both bind equally
strong. Notably, adding one phenyl function yields a smaller HPLC–MS measurements
increase in binding affinity than adding a second one, pointing The HPLC setup was based on a Shimadzu system and con-
towards a cooperative contribution of both phenyl rings. Pure F sisted of a CBM-20A controller, a SIL-HTA auto sampler, two
is the weakest tested binder, which hints at the selectivity 10ADVP pumps, a DGU-14A degasser, a CTO-10AVP column
towards dipeptides.
oven and an SPD-10AVP variable wavelength detector. The
1592